Review on 6G-oriented terahertz communication channel

With the large-scale commercialization of 5G mobile communication systems in the world, 6G technology is becoming a new hotspot of global information technology research. The terahertz band (300GHz to 3THz) has extremely rich frequency resources, which can support ultra-high speed wireless communication from 100Gbps to 1Tbps, improve the existing 5G peak transmission rate by one or two orders of magnitude, and meet the requirements of new applications such as ultra-high resolution holographic communication and meta-universe. An in-depth understanding of the propagation channel is crucial to the design of the communication system. This paper summarizes the research status of the terahertz channel, including its characteristics in the atmosphere, the current channel propagation model, and terahertz channel measurement. At the same time, the future research direction of terahertz communication is analyzed and prospected, and the possible challenges in the future research are discussed, which will help promote the application of terahertz communication technology to 6G communication. Due to technical limitations, terahertz channels have not been fully developed and utilized in the field of communication, and there are many technical bottlenecks that cannot be broken. In the future, new channel models may be built, and there will be opportunities for AI-assisted channel measurement.


Introduction
A new generation of wireless communication systems will be produced every two decades, starting with the first generation of mobile communication technology (1G) in the 1980s and ending with the fifth generation of mobile communication technology (5G) today.With the continuous upgrading of mobile communication technology, wireless network performance such as data transmission rate, communication quality, and network delay, will also continue to improve.Terahertz communication is generally regarded as the core component of the future 6G mobile communication system.China and other advanced countries in the field of mobile communication, such as the United States, Europe, and Japan, are making efforts to implement 6G technology.According to the Horizon Europe plan, 6G terahertz communication technology was developed in the European Union in 2017.In 2019, the Federal Communications Commission (FCC) of the United States opened the 95 GHz to 3 THz frequency band for future mobile communication applications, and formulated relevant plans for the construction of a terahertz broadband wireless communication backbone network.This study makes a detailed summary of the research status of terahertz, including Terahertz channel characteristics in the atmosphere, terahertz channel propagation model and terahertz channel measurement.This paper looks forward to the feasible direction of terahertz communication in the future and enriches the understanding of terahertz channel, which is very helpful for the future terahertz channel modeling and the design of communication systems.

Terahertz atmospheric absorption attenuation
The absorption attenuation of atmospheric molecules is caused by the absorption of electromagnetic wave energy by molecules in the atmosphere.Part of the excited molecules in the atmospheric molecules vibrate at a certain frequency and resonate with the electromagnetic wave.Part of the electrons undergo energy level transitions, and the energy of the electromagnetic wave is transformed into the kinetic energy of the molecules, that is, the energy of the electromagnetic wave is lost [1].
To determine THz atmospheric absorption and attenuation, International Radio Consultative Committee (CCIR) suggests using the condensed formula for atmospheric absorption at sea level.Water vapor density, temperature, and pressure are required as inputs for the THz atmospheric absorption attenuation model, according the International Telecommunication Union Radiocommunication Sector (ITU-R) P.676-9 recommendation.The absorption loss of atmospheric molecular absorption attenuation (unit: dB/km) can be expressed as formula (1).
Where   is the absorption attenuation of oxygen molecules under dry air conditions,   is the characteristic attenuation of water vapor molecules in the atmospheric environment.Both values are related to frequency , in dB/km.

Terahertz atmospheric scattering attenuation
In order to use terahertz electromagnetic waves in future 6G communication, it is crucial to contemplate the scattering effect of meteorological particles such as fog, rain and snow in the actual communication environment.When a terahertz electromagnetic wave meets suspended particles during transmission, its energy will be scattered outside the propagation path, resulting in attenuation of the wave energy.
Fog, also known as clouds blocked by the ground, has a similar micro-physical structure to clouds and is found in areas with high water vapor content, such as mountains and beaches.According to Recommendation ITU-R P.840-7, the Rayleigh approximation is used to calculate cloud attenuation for particles of such small size as clouds.
(, ) =   (, ) (2) Where   (, ) is the attenuation coefficient of the liquid water ratio;  is the density of liquid water;  is the operating frequency, in GHz;  is the temperature of liquid water, in .
In a medium that contains raindrops, the attenuation of terahertz electromagnetic wave propagation is proportional to the amount of precipitation.In Recommendation ITU-R P.838-3, for a given precipitation rate R (unit: mm/h), no matter whether the rainfall is large or small, the attenuation caused by rainfall will increase exponentially.This attenuation model (unit: dB/km) can be expressed as formula (3)   = ()   (3) where  and  is a function of operating frequencies from 1 GHz to 1 THz, which are also affected by parameters such as temperature, polarization mode, and height.
There is no relevant recommendation of ITU-R to provide a model for predicting snow failure because there is little research on the path loss caused by snowfall.Since the size of snowflakes is equivalent to the wavelength of terahertz, F. Norouziari et al. point out that, under the same precipitation rate, the attenuation caused by snowfall is 3 times that of rainfall [2].

Deterministic channel modeling
According to current application scenarios, a wireless channel model is created using the deterministic channel modeling method based on an analysis of optical and electromagnetic propagation theory.Its benefits include not requiring real measurement, but its drawbacks include the need for extremely specific application scenario data and considerable computational complexity [3].In particular, two typical approaches are Ray Tracing (RT) and Finite Difference Time Domain (FDTD) for deterministic channel modeling.
The RT approach uses geometric optics and the high-frequency approximation of Maxwell's equations to simulate the movement of electromagnetic waves.To be more specific, the ray tracing method models the channel between the transceiver based on geometric optics, geometric diffraction theory, uniform diffraction theory, and Kirchhoff theory after first determining the position of the transceiver [4].
The FDTD method is a numerical analysis method that can solves Maxwell's equations directly.The magnetic and electric fields are alternately sampled in the time and spatial domains to achieve FDTD.This ensures that each magnetic field sampling point is surrounded by six electric field sampling points, and vice versa.Through the central difference method, the Maxwell equation is discretized for later calculation [5].

Statistical channel modeling
Statistical channel modeling uses a measurement platform to conduct actual measurements in application scenarios, adjusting actual data to obtain the empirical distribution and statistical characteristics of each channel parameter.Finally, the channel is reconstructed based on statistical characteristics.
Non Geometric Stochastic Channel Model (NGSM), known as the Parameter Stochastic Model (PSM), is a purely random model.Regardless of the actual propagation environment, they establish a probability distribution function to characterize and calculate parameters like the Direction of Departure (DoD), Direction of Arrival (DoA), and latency.Because of its straightforward structure and cheap computing complexity, NGSM is widely utilized in channel modeling, according to Li Yuanbo [4].However, it is difficult for them to describe the complex relationship between parameters, especially spatial consistency, and the relationship between DoD, DoA, and time variation of delay as the device moves farther.

Channel sounders
Terahertz channel measurement can be roughly divided as two methods: time range measurement and frequency range measurement, corresponding to two types of channel sounders.The time range channel sounder is constructed using the Time Domain Correlation (TDC) or the Time Domain Spectroscopy (TDS) methods.On the other hand, the frequency range channel sounder is constructed utilizing a Vector Network Analyzer (VNA) [6].

Channel measurement based on TDC method
Measurement based on the TDC method is not only widely used for indoor activities, but also for many outdoor scenes.According to A. Molisch, the TDC-based sounders are based on the idea of transmitting signals from the transmitters [7].The autocorrelation of the signal is an approximation of the Dirac Delta function [6].Therefore, if the signal () is transmitted through a channel with impulse response ℎ(, ), the received signal � ′ , �at a specific instant of time  ′ can be obtained by formula (4).
A hybrid channel identification method that combines correlation modes and real-time spectrum diffusion modes was proposed by George R. MacCartney and T. Rapport [8].The detection supports propagation delay measurement and multi-path component identification, and the delay resolution of each mode is 2ns.The correlation mode specifically supports measuring long-distance parameters, such as deceleration and angular dispersion, with a loss of propagation path up to 185dB.On the other hand, the spread spectrum mode is more suitable for measuring short-range and small-scale time parameters, as well as the Doppler body block with a dynamic range of 40dB.

Channel measurement based on TDS method
According to Anirban Ghosh and Minseok Kim, using a TDS-based channel sounder for channel measurement has many advantages, such as a large extended bandwidth and a high channel acquisition speed [9].However, this kind of sounder also has the disadvantage of low power output, which makes its measuring distance very limited.To sum up, the sounder based on the TDS method is the first choice for studying the reflection properties of different materials, scattering through objects in regular environments, developing interference models with high node density, or studying the effects of weather on THz communication.
For purpose of correctly model the propagation environment to support indoor terahertz microcellular communication, it is essential to have a well-rounded understanding of the impact of reflectivity on conventional materials used in a ordinary indoor circumstance.In order to investigate the impacts of frequency-dependent reflectivity and incident angle of layered building materials like white paint and gypsum on double-glazed windows with frequencies between 100 and 500 GHz, Christian Jansen et al. used a TDS-based sounder [9].

Channel measurement based on VNA
VNA is the sub-channel narrowband's contrast transfer function (CTF), which is sequentially measured in the frequency domain.The CTF of the entire communication bandwidth is the collection of all subnarrowband CTFs; Then carry out inverse fast Fourier transform (IFFT) to obtain the corresponding channel impulse response CIR [10].
Due to the simple operation and low learning cost of VNA, VNA-based methods are widely used in terahertz channel measurement.However, this method also has some disadvantages, such as low output power.Moreover, it usually takes a long time for a frequency-sweep, which limits its use in dynamic scenarios.Furthermore, transmitter (Tx) and receiver (Rx) must be connected to the identical VNA by cable.The high attenuation of the THz signal in the cable significantly limits the distance between Tx and Rx.Therefore, voice equipment based on VNA THz channel is usually only applicable to shortrange static THz channel [11].Yi Haofan's team proposed a two-port calibration method: the short−open−load−thru (SOLT) calibration method, which is used for the whole measurement system, including the waveguide, spread spectrum module, cable, and VNA itself [10].This method is widely used, easy to master, suitable for most applications, and can provide excellent accuracy and repeatability.

Conclusion
This paper summarizes the current research from three aspects: terahertz channel characteristics, modeling and measurement, and concludes that terahertz atmospheric absorption attenuation and atmospheric scattering attenuation under three scenarios of fog, rain, and snow will have a significant impact on terahertz communication, and the resulting losses can be calculated by the corresponding formula.The three channel modeling methods RT, FDTD, and NGSM, are suitable for real-life scenarios and have potential to be applied to future 6G mobile communications.Finally, most of the channel measurements used in existing research work are based on TDC, TDS methods, or VNA and are all done in a static manner.This paper does not propose a better terahertz channel modeling and measurement method, which has certain shortcomings.Looking ahead to future terahertz channel research, several challenging issues await resolution.Firstly, the research on path loss caused by snowfall is relatively scarce.Secondly, whether a new channel model is needed.In the meantime, it may be possible to consider using artificial intelligence to accomplish channel measurement and establish new channel sounders in the future.